I still love Twitter. A lot of us still love Twitter. But it’s past time to admit it’s an abusive relationship. (“Yes, he hits me sometimes, but it’s only for the retweets.”)
The common wisdom is that the Big Blue Bird’s problem is their lack of moderation, that the service is Exhibit A in the case against Silicon Valley’s belief that you can solve everything with algorithms. I think that’s some of it, but I don’t think it’s all of it. When your software becomes global community infrastructure, the choices reflected in your design have profound effects on behavior. It’s a choice, for instance, to offer no privacy controls other than “protecting” your account. That one choice alone is a large part of why Twitter is so hospitable to harassers: your only option for controlling who engages with you is flipping your entire feed between open and locked down, and—given that anyone you follow can inject anything into your timeline via retweet—aggressively curating not just who you follow, but who you allow retweets from.
Here are some other choices Twitter’s made. “Favorites” are public accolades, not private bookmarks. Mechanisms for retweets and quote tweets are baked in. Official clients stream notifications about not just who favorited and retweeted you, but who favorited and retweeted your retweets. And let’s not even get into who gets verified and what verification offers. None of these choices are necessarily wrong in either a technical or moral sense. But they’ve created a culture that rewards painting everything in the starkest, loudest terms possible.
There’s a metric crapton of political tweets across the partisan spectrum that I could point to, but as I was writing this piece, a bag of “Lady Doritos” dropped into my lap.
PepsiCo CEO Indra Nooyi gave an interview to the Freakonomics podcast in which she observed that women ate Doritos differently than men did (“they don’t like to crunch too loudly in public”) and said the company was getting ready to launch “snacks for women that can be designed and packaged differently.” The Sun, a UK tabloid, reported this as “Doritos to launch crisps for WOMEN because they don’t like crunching loudly or licking their fingers, boss reveals.” This led to a veritable tortillanado of hot take tweets about snack food sexism.
But wait! Then came the New York Times reporting “Not a Real Thing, Company New Says,” which quoted PepsiCo’s gently acerbic retort, “We already have Doritos for women. They’re called Doritos.” Snap! Fake news! Well, yes and no. The quotes from Ms. Nooyi in the last paragraph are true; Frito-Lay is working on “snacks for women,” whatever the hell that may mean. The fake news part—in the sense that the Sun came up with it, not Nooyi—was the existence of “Lady Doritos.” Gosh, what an outrage-inducing, easily hashtaggable name they invented! Surely that couldn’t have been their intent. Ha. Ha ha. As of this writing, we’re 48 hours into Chipghazi, and the Twitter trends are just starting to ebb.
And this is a problem inherent in Twitter’s design that may not be solvable. Even if Twitter engineers could just go into the database and type
DELETE FROM users WHERE is_nazi = 1, the software’s literally designed to reward superficial hot takes. It’s optimized for tweets that make you go “yeah, get those fuckers!” rather than tweets that make you go “hmm.”
When was the last time you scrolled through your Twitter timeline and felt smarter, happier, and generally more at peace with the world?
Mastodon and Micro.blog both propose that the solution to Twitter’s ills is decentralization. Mastodon has multiple “instances,” like Twitter servers, that each have their own rules and community guidelines. Because all the instances can interact with one another, you can follow any Mastodon user, not just the ones on your instance. Micro.blog is, if anything, more radical: a set of open standards that let good old fashioned weblogs interact with one another in Twitter-esque fashion. You can use it just like Twitter, but under the hood it’s using an RSS-like system to build your timeline. They can host your own (paid) micro.blog, which is a full-featured Jekyll install under the hood, but you could host your own blog wherever and on whatever software you want.
So far, these solutions are working, but I’m worried that—particularly in Mastodon’s case—it’s not because they’ve chosen a more resilient design, it’s simply because the community is so much smaller. There’s less social reward for turning the volume on everything up to 11 when the audience is tiny. But Mastodon makes many of the same choices Twitter has, including favorites, quotes (“embeds”), and retweets (“boosts”), then stirs in the questionable belief that moderation issues are effectively moot under their federated server model.
Micro.blog deliberately has no retweet mechanism. Favorites are just private bookmarks. As far as I can tell you can’t even get a list of followers. Unlike Mastodon, Micro.blog shows replies people make to people you aren’t following, the way Twitter did in its first couple of years. All this adds up to a surprisingly friendly, conversational timeline. (Also, Micro.blog’s first hire has been a community manager, which says a lot about their philosophy here.) But as I alluded above, if you want to use it just like Twitter—i.e., no work on your part—you need to pay them to host your blog. They’re looking at it as a turnkey blog hosting service, but if it’s perceived as “like Twitter but with less features for $5 a month,” that’s a problem.
Yet both Micro.blog and Mastodon are just…nicer. I think Micro.blog is the better of the two, in no small part for the UX choices they’ve made that are explicitly the opposite of both Twitter and Mastodon, but Mastodon’s free nature gives it the potential to grow further. Either way, though, both of them have one huge advantage: they’ve seen the shitshow that’s turned Twitter into a Dead Bird Walking, and they can say, “You know what? Let’s not do that.”